Price-based measures to reduce
alcohol consumption
Andrew Leicester (with Rachel Griffith and Martin O’Connell)
Health Economics Research Unit, University of Aberdeen
12th March 2013
© Institute for Fiscal Studies
Introduction
• Governments long been concerned about excess alcohol intake
• Excise taxes have been main price-based intervention to date
– annual real-terms increases since 2008, planned to 2014
• Recent consideration of alternative policies:
1. Minimum unit price (MUP)
– floor price based on alcohol content
– 1 unit = 10ml alcohol
2. Banning quantity-based alcohol promotions
– prohibit multi-buy deals, discounts on multi-packs
© Institute for Fiscal Studies
Current analysis: preliminary findings
• We compare the effectiveness of different price-based policies:
– MUP of 45p
– quantity discount ban
– reformed excise tax system targeting alcohol strength
• We use detailed longitudinal household off-trade purchase data
• Make simple assumptions about behavioural responses and ask:
– are policies well-targeted on ‘problem’ drinkers?
– what are the implications for spending, tax revenue, firm revenue?
• Our results so far suggest that:
– tax reform would be better-targeted than MUP
– tax reform also raises revenue for government, not industry
– quantity discount bans are very poorly-targeted
© Institute for Fiscal Studies
Outline structure
• Background
– data on alcohol intake, prices and affordability
– brief recap on economic rationale for different reforms
• Methods
– dataset for analysis
– descriptive figures for off-trade alcohol purchasing behaviour
– description of policy reforms analysed
– assumptions about demand- and supply-side responses
• Results
– descriptive analysis of quantity discount ban
– comparative analysis of effect of MUP and excise tax reform
• Conclusions, limitations and future plans
© Institute for Fiscal Studies
Outline structure
• Background
– data on alcohol intake, prices and affordability
– brief recap on economic rationale for different reforms
• Methods
– dataset for analysis
– descriptive evidence on alcohol purchasing behaviour
– description of policy reforms analysed
– assumptions about demand- and supply-side responses
• Results
– descriptive analysis of quantity discount ban
– comparative analysis of effect of MUP and excise tax reform
• Conclusions, limitations and future plans
© Institute for Fiscal Studies
Intake trend differs in UK from near-neighbours
Average intake of alcohol (litres per adult), 1970 to 2010
UK
20
France
15
Germany
10
Italy
5
Spain
Source: OECD Health Statistics 2012
© Institute for Fiscal Studies
2010
2006
2002
1998
1994
1990
1986
1982
1978
1974
0
1970
Annual alcohol intake (litres p.c.)
25
Real alcohol price flat, but on/off differential
Relative alcohol price indices,1990 to 2012
Beer (on)
130
120
Beer (off)
110
100
Wine &
spirits (on)
90
80
Source: Calculated from ONS Retail Prices Index data
© Institute for Fiscal Studies
2012
2010
2008
2006
2004
2002
2000
1998
1996
All alcohol
1994
60
1992
70
Wine &
spirits (off)
1990
Real index (Jan 1990=100)
140
Alcohol affordability has recently fallen back
130
125
120
115
110
105
100
95
90
85
80
Earnings
Price
Source: Calculated from ONS Retail Prices Index data and average earnings data.
Note: average earnings and affordability indices are 12-month moving averages.
© Institute for Fiscal Studies
2012
2010
2008
2006
2004
2002
2000
1998
1996
1994
1992
Affordability
1990
Real index (Jan 1990=100)
Real alcohol prices and real incomes, 1990 to 2012
Outline structure
• Background
– data on alcohol intake, prices and affordability
– brief recap on economic rationale for different reforms
• Methods
– dataset for analysis
– descriptive evidence on alcohol purchasing behaviour
– description of policy reforms analysed
– assumptions about demand- and supply-side responses
• Results
– descriptive analysis of quantity discount ban
– comparative analysis of effect of MUP and excise tax reform
• Conclusions, limitations and future plans
© Institute for Fiscal Studies
Rationale for price-based intervention
• Excess alcohol intake generates external costs
– crime, anti-social behaviour, family problems
– public costs of treating health harms, productivity spillovers
– time inconsistency could generate externalities for future self
• Effective interventions depend on:
– targeting ‘problem’ drinking: price according to marginal externality
• varies across drinkers and drinks
• may not be able to target high cost consumption directly
– response to price changes
• does demand for high cost units fall as price rises?
• how do firms respond to intervention?
• Optimal policy needs to trade-off benefits from reducing problem
consumption against costs of reducing low-cost consumption
© Institute for Fiscal Studies
Our analysis
• We explore whether policies effectively target heavy drinkers
– average weekly intake above recommended levels
– drinking small amounts unlikely to generate external costs
• Other drinking habits may also be of policy concern
– binge drinking, under-age drinking
– we do not have data on this specifically but interesting to explore
• Different reforms effectively targeted if heavy drinkers tend to:
– buy cheap alcohol (minimum pricing)
– buy strong alcohol (tax reform)
assess with data
– buy in bulk (quantity discount ban)
• We do not directly estimate price responsiveness to policy
– assume values drawing on literature
© Institute for Fiscal Studies
Outline structure
• Background
– data on alcohol intake, prices and affordability
– brief recap on economic rationale for different reforms
• Methods
– dataset for analysis
– descriptive evidence on alcohol purchasing behaviour
– description of policy reforms analysed
– assumptions about demand- and supply-side responses
• Results
– descriptive analysis of quantity discount ban
– comparative analysis of effect of MUP and excise tax reform
• Conclusions, limitations and future plans
© Institute for Fiscal Studies
Data
• Market research in home scanner data from Kantar Worldpanel
– GB household-level panel, sample size c. 25,000
– households can drop out at any time, mean duration c. 2 years
– demographics recorded at sign up, update approx. every 9 months
• Records each item purchased on each trip at barcode level
– price matched in from till receipts, includes deal information
– calculate ABV % from data, online sources and ONS
– use HMRC data to match in excise tax levied on each purchase
• Use 52 weeks from Nov 2009 to Oct 2010
– select sample of households who report spending consistently
– 21,542 hhs (median: 301 day duration), 522,125 alcohol purchases
• Calculate units purchased per adult per week for each household
© Institute for Fiscal Studies
Data
• Data weighted so that observed spending in sample matches
total off-trade spending from ONS National Accounts
– lines up well with tax and quantity data from HMRC by alcohol type
• Key strengths
– long observation period: purchases should approximate intake
– large sample size
– detailed purchase and price information
• Key weaknesses
– no on-trade purchases
– household-level not individual-level data
– purchase not consumption (cannot measure e.g. bingeing)
© Institute for Fiscal Studies
Outline structure
• Background
– data on alcohol intake, prices and affordability
– brief recap on economic rationale for different reforms
• Methods
– dataset for analysis
– descriptive evidence on alcohol purchasing behaviour
– description of policy reforms analysed
– assumptions about demand- and supply-side responses
• Results
– descriptive analysis of quantity discount ban
– comparative analysis of effect of MUP and excise tax reform
• Conclusions, limitations and future plans
© Institute for Fiscal Studies
Alcohol purchase behaviour by intake level
Grouped by average off-trade units per adult per week
N
%
Litres of
drink per
adult/week
≤7
16,427
76%
0.2
10.5%
45.4
17.0%
>7, ≤14
2,586
12%
1.1
11.0%
42.5
16.1%
>14, ≤21
1,103
5%
1.8
12.3%
41.4
14.5%
>21, ≤35
898
4%
2.7
12.9%
39.5
13.7%
>35
528
2%
4.9
14.3%
37.5
11.5%
Units
Avg.
strength
(ABV)
Avg. pence
per unit
% units on
quantity
discount
• NHS guidelines for ‘hazardous’ drinking:
– 14+ units per week (6+ in one session) for women
– 21+ units per week (8+ in one session) for men
• Our figures suggest 6 to 11% of households ‘hazardous’ on basis
of average off-trade consumption
© Institute for Fiscal Studies
Outline structure
• Background
– data on alcohol intake, prices and affordability
– brief recap on economic rationale for different reforms
• Methods
– dataset for analysis
– descriptive evidence on alcohol purchasing behaviour
– description of policy reforms analysed
– assumptions about demand- and supply-side responses
• Results
– descriptive analysis of quantity discount ban
– comparative analysis of effect of MUP and excise tax reform
• Conclusions, limitations and future plans
© Institute for Fiscal Studies
Minimum unit price
• Home Office has recommended a 45p MUP for England & Wales
– policy currently under consultation
• Scottish government has legislated for a 50p MUP
– implementation delayed pending a legal challenge
• We assess the impact of a 45p MUP across Britain
– assume the 45p rate was applicable in April 2012
– assume rate uprated each year in line with excise duties
– thus we actually simulate rates of 38p and 40p
– change applies on 29 March 2010 when excise taxes rose
© Institute for Fiscal Studies
MUP affects majority of off-trade units
Proportion of units below simulated MUP by alcohol type, 2010
90
% of units below MUP
80
70
60
50
40
30
20
10
0
Avg. pence per unit
Cider
Spirits
Beer
Wine
Alcopops
All
28.1
40.6
38.7
45.5
81.2
51.4
Source: Calculated from Kantar Worldpanel 2010 data. MUP applied is 38p, rising to 40p from 29/3/10.
© Institute for Fiscal Studies
Reforms to alcohol excise taxes
• Excise taxes are specific, vary by alcohol type and strength
– raise around £10 billion, just under 2% of receipts
• Broad structure governed by EU Directive (92/83/EEC)
– beer, spirits taxed by alcohol content
– wine, cider taxed by volume of product (within strength bands)
© Institute for Fiscal Studies
Baseline alcohol excise tax structure, 2009/10
Excise duty per unit of alcohol (pence)
Excise duty rate per alcohol unit, by type and strength
60
Wine
Sparkling wine
50
Spirits and alcopops
Beer
40
Cider
30
20
10
0
0
4
8
12
16
20
24
% ABV strength
Source: Calculated from HMRC data.
© Institute for Fiscal Studies
28
32
36
40
Reforms to alcohol excise taxes
• Taxes do not look well-targeted on strong alcohol
• We consider more radical reform in which tax is levied on alcohol
content directly for all alcohol types
• Examine three possibilities:
– tax all units of alcohol equally
– vary rate by type but not strength
– vary rate by type, rate increases with strength
Note: These rates applied before April 2010. We assume rates then rise in line with actual duties.
© Institute for Fiscal Studies
Reformed alcohol tax structures, 2009/10
Wine
Sparkling wine
Spirits and alcopops
Beer
Cider
60
50
40
30
Excise duty/unit (pence)
Excise duty/unit (pence)
Simulated excise duty rates per alcohol unit, by type and strength
20
10
0
60
50
40
30
20
10
0
0
4
8
12
16
20
24
28
32
36
40
0
4
8
12
Excise duty/unit (pence)
60
Wines and spirits
50
20
24
28
32
36
40
% ABV strength
Beers, ciders and alcopops
40
30
20
10
Excise duty/unit (pence)
% ABV strength
16
60
Wines and spirits
50
Beers, ciders and alcopops
40
30
20
10
0
0
0
4
© Institute for Fiscal Studies
8
12
16 20 24
% ABV strength
28
32
36
40
0
4
8
12
16
20
24
% ABV strength
28
32
36
40
Reforms to alcohol excise taxes
• Taxes do not look well-targeted on strong alcohol
• We consider more radical reform in which tax is levied on alcohol
content directly for all alcohol types
• Examine three possibilities:
– tax all units of alcohol equally
– vary rate by type but not strength
– vary rate by type, rate increases with strength
• The particular rates we choose are designed to achieve same
aggregate reduction in units as the MUP simulation
– impact across households and on revenues will differ
• Aim is not to suggest these particular reforms are ‘optimal’
– want to compare to MUP on basis they achieve a common outcome
© Institute for Fiscal Studies
Quantity discount ban
• Implemented in Scotland (off-trade) in October 2011
• Under consultation in England and Wales
• Policy applies to:
– quantity-based special offers (BOGOF, 3F2, 5% off 6 bottles, etc.)
– bulk discount of given container size/brand combination (single
500ml can = £1, 24-pack ≥ £24) if smaller pack available in-store
• Policy does not apply to:
– price-based special offers (50% off), extra free offers
– bulk discount across container sizes (2-litre bottle costs less than
21-litre bottle)
• We describe importance of quantity discounts and large multipacks in alcohol purchases of each household
© Institute for Fiscal Studies
Outline structure
• Background
– data on alcohol intake, prices and affordability
– brief recap on economic rationale for different reforms
• Methods
– dataset for analysis
– descriptive evidence on alcohol purchasing behaviour
– description of policy reforms analysed
– assumptions about demand- and supply-side responses
• Results
– descriptive analysis of quantity discount ban
– comparative analysis of effect of MUP and excise tax reform
• Conclusions, limitations and future plans
© Institute for Fiscal Studies
Assumed responses to policy interventions
• Demand-side
– assume all consumers have alcohol price elasticity of -0.5
• central figure in Wagenaar et al. (2009) meta-analysis of 112 studies
– calculate policy-induced change in avg. price for each household
– apply elasticity at the household level
– no wider demand-side responses
• e.g. cross price effects between alcohol and other consumption
• Supply side
– assume firms are essentially passive
– changes in excise tax are fully passed into final prices
– prices below MUP are raised to threshold, no other price changes
– no impact on product range, advertising and promotion etc.
© Institute for Fiscal Studies
Outline structure
• Background
– data on alcohol intake, prices and affordability
– brief recap on economic rationale for different reforms
• Methods
– dataset for analysis
– descriptive evidence on alcohol purchasing behaviour
– description of policy reforms analysed
– assumptions about demand- and supply-side responses
• Results
– descriptive analysis of quantity discount ban
– comparative analysis of effect of MUP and excise tax reform
• Conclusions, limitations and future plans
© Institute for Fiscal Studies
Multibuy deals in off-trade alcohol (% units)
By purchase level and alcohol type, 2010 data
Beer
Cider
Wine
Fabs
Spirits
All
≤ 7 units
25.9
22.8
16.1
23.8
4.5
17.0
≤ 14 units
24.6
18.7
17.1
23.6
4.4
16.1
≤ 21 units
23.4
17.9
16.2
18.7
3.5
14.5
≤ 35 units
21.9
17.5
16.3
19.1
3.3
13.7
> 35 units
17.7
22.1
14.4
14.5
2.3
11.5
All
23.4
20.2
16.0
22.0
3.5
14.6
Source: Calculated from Kantar Worldpanel
© Institute for Fiscal Studies
Multipacks in off-trade alcohol
Average ‘items per pack’, by purchase level and alcohol type, 2010 data
Beer
Cider
Wine
Fabs
Spirits
≤ 7 units
6.1
2.5
1.0
1.9
1.0
≤ 14 units
6.8
2.8
1.0
1.7
1.0
≤ 21 units
6.9
2.9
1.0
1.8
1.0
≤ 35 units
7.2
2.8
1.0
1.7
1.0
> 35 units
6.2
2.5
1.0
1.6
1.0
All
6.5
2.7
1.0
1.8
1.0
Source: Calculated from Kantar Worldpanel
© Institute for Fiscal Studies
Typical ‘item size’ in off-trade alcohol
Avg. bottle/can size (mls), by purchase level and alcohol type, 2010 data
Beer
Cider
Wine
Fabs
Spirits
≤ 7 units
437
831
746
458
719
≤ 14 units
443
1,077
788
495
762
≤ 21 units
452
1,171
825
427
768
≤ 35 units
449
1,409
874
457
795
> 35 units
455
1,513
1,000
457
853
All
445
1,126
828
460
778
Source: Calculated from Kantar Worldpanel
© Institute for Fiscal Studies
Quantity discount ban: key points
• Policy seems badly-targeted:
– heavy drinkers do not rely more heavily use of multibuys
• have not yet looked at variation in type or ‘size’ of offers used
– heavy drinkers do not buy larger multi-packs
– heavy drinkers buy larger containers, but not affected by policy
• Supply-side response means policy need not raise prices:
– increase use of price discounts (£3.33 per bottle, not 3 for £10)
– reduce price of small packages
• both allow people to obtain lower unit price at lower quantity – could be beneficial
– remove small packages from sale
• limit ability to obtain small quantities – could be harmful
© Institute for Fiscal Studies
Outline structure
• Background
– data on alcohol intake, prices and affordability
– brief recap on economic rationale for different reforms
• Methods
– dataset for analysis
– descriptive evidence on alcohol purchasing behaviour
– description of policy reforms analysed
– assumptions about demand- and supply-side responses
• Results
– descriptive analysis of quantity discount ban
– comparative analysis of effect of MUP and excise tax reform
• Conclusions, limitations and future plans
© Institute for Fiscal Studies
Consumption response to reforms
Average of household-level responses, by intake group
Units per adult per week
Avg. consumption change (%)
≤7
≤ 14
≤ 21
≤ 35
>35
0
-1
-2
-3
-4
-5
-6
-7
-8
-9
-10
MUP
Type-varying tax
Single tax rate
Type- and strength-varying tax
Source: Calculated from Kantar Worldpanel. Note: excludes abstainers.
© Institute for Fiscal Studies
Distributional issues
Effect on household grocery budgets, by income and intake group
% increase in grocery budget
0.5
MUP
Type-strength tax
0.4
0.3
0.2
0.1
0
< £10k
< £20k
< £30k
< £40k
< £50k
< £60k
Gross annual household income group
Source: Calculated from Kantar Worldpanel. Note: excludes abstainers.
© Institute for Fiscal Studies
≥ £60k
Aggregate effects
Baseline
Off-trade
units (bn)
Alcohol
spending (£ bn)
Tax revenue
duty+VAT (£ bn)
Firm revenue
(£ bn)
37.04
15.33
8.95
6.39
-2.40
+0.55
-0.29
+0.84
(-6.5%)
(+3.6%)
(-3.3%)
(+13.2%)
-2.39
+0.65
+1.04
-0.39
(-6.5%)
(+4.3%)
(+11.6%)
(-6.0%)
-2.41
+0.75
+1.15
-0.40
(-6.5%)
(+4.9%)
(+12.9%)
(-6.3%)
-2.40
+0.64
+0.98
-0.34
(-6.5%)
(+4.2%)
(+11.0%)
(-5.4%)
Change due to policy reform:
MUP
Single tax rate
Type-varying
tax
Type-strength
tax
Source: Calculated from Kantar Worldpanel.
© Institute for Fiscal Studies
Summary of policy impacts
• MUP equivalent to 45p in 2012 prices
– average impact roughly 2bigger for heavy drinkers than light
– mildly regressive
– generates windfall of around £840m for industry
– loss of around £290m in tax receipts (duty -£370m, VAT +£80m)
• Excise tax reform targeting stronger alcohol
– average impact roughly 3bigger for heavy drinkers than light
– mildly regressive
– reduces industry revenue by around £340 million
– generates around £980m in tax receipts (duty +£890m, VAT +£90m)
© Institute for Fiscal Studies
Outline structure
• Background
– data on alcohol intake, prices and affordability
– brief recap on economic rationale for different reforms
• Methods
– dataset for analysis
– descriptive evidence on alcohol purchasing behaviour
– description of policy reforms analysed
– assumptions about demand- and supply-side responses
• Results
– descriptive analysis of quantity discount ban
– comparative analysis of effect of MUP and excise tax reform
• Conclusions, limitations and future plans
© Institute for Fiscal Studies
Conclusions
• Quantity discount ban poorly targeted and may well have limited
price effect – does not seem like a good policy
• MUP often seen as more targeted than increase in excise duty
• But reform of duty targeting strong alcohol even better-targeted
– and would raise tax revenue rather than give industry a windfall
• European dimension important …
– EU Directive prohibits reform of alcohol taxes as we suggest
– government should lobby for necessary reform to allow tax change
– MUP legality under EU Directives?
© Institute for Fiscal Studies
Further analysis
• Allow elasticity to vary across alcohol types and intake level
– draw on literature and meta-analysis
• relatively consistent variation by type (e.g. beer more elastic)
• less clear how elasticities vary across intake group
– estimate off-trade demand model using Worldpanel data
• Relax assumption of passive firm responses
– allow firms to re-optimise
• wider price responses, product choices, advertising and promotion strategies
– make particular assumptions or try to model explicitly
© Institute for Fiscal Studies